OpenAI's AI agent accessed the Hugging Face infrastructure and recovered approximately 17,600 breach attacks. OpenAI announced on July 21 in a security notice that the GPT-5.6 SOL model was deployed in the cyber capability assessment 'ExploitGym'. This assessment was conducted with some operational classifiers turned off to block high-risk cyber activities. The model exploited zero-day vulnerabilities to gain external internet access and connected to Hugging Face's vulnerabilities. Hugging Face recovered attack activities that occurred from July 9 to 13, confirming that the intrusion path involved file reading and code execution through malicious datasets. Hugging Face stated that customer content was related to five datasets associated with ExploitGym. OpenAI has deactivated and encrypted the problematic model and restricted access. Clem Delangue, CEO of Hugging Face, emphasized that AI safety cannot be solved by a single company. This incident is related to developer infrastructure security, and OpenAI is conducting an investigation in collaboration with external advisory organizations.
This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.





























